Researchers Create AI Algorithm To Predict Heart Attack Risks

May, 2022 - By SMI

The new artificial intelligence algorithm measures build-up of plaque in coronary arteries accurately and predicts risk for patients to suffer a heart attack within five years.

Cardiovascular diseases are one of the foremost reasons of deaths all over the world. Now a team of researchers at Cedars-Sinai Medical Center developed a novel AI tool that can measure plaque build-up in coronary arteries accurately and it can also predict risk for heart attacks within five years for patients. The new tool is considered to be a promising method for prevention of heart attacks in minimum time. As of now, the tool is still to be validated to bring it into clinical use.

Currently, computed tomography angiography (CTA) is most common method used for evaluating patients with heart disease and CTA imaging of plaque build-up in heart is an efficient way to predict risks of heart attack of patients. Measuring this plaque build-up in coronary arteries takes around 30 minutes. However, with the new AI tool it can be done in seconds. In this new research, the team created this new AI tool by training an algorithm to identify deposits of plaque with the use of a dataset of CTA images from 921 patients. Then, the team validated the tool using a test set of CTA images from hundreds of patients, which delivered results that were close to readings of human experts.

Furthermore, the team set a number thresholds of plaque volume for the tool to predict risks of heart attacks in future. The team found that the new AI tool was able to stratify patients into two groups, one with high risk and one with low risk of heart attacks within five years after CTA imaging. The technology is still in its early stages to bring it into clinical use, however, this study demonstrated the future of AI in medicine.